Next Publication In:
Days: 00
Hours: 00
Minutes: 00
Seconds: 00

A HYBRID FINITE-STATE AND RULE-BASED ARCHITECTURE FOR THE ANGIKA MORPHOLOGICAL ANALYZER GENERATOR

📘 Volume 11 📄 Issue 1 📅 January 2026

👤 Authors

Alok Kumar, Manisha Kumari Deep 1
1. 1. Department of Computer Science Engineering, YBN University Ranchi, Department of Computer Science Engineering, 2.School of Computer Science and IT, YBN University Ranchi

📄 Abstract

Morphological analysis is a core component of natural language processing, particularly for mor- phologically rich and low-resource languages where data-driven approaches are often impractical. This paper presents a hybrid finite-state and rule-based morphological analyzer generator for Angika, an under-resourced Indo-Aryan language. The proposed system adopts a generator-oriented architecture that separates linguistic knowledge from core processing mechanisms, enabling modularity, extensibil- ity, and reuse across related languages. Finite-state transducers are employed to model regular inflec- tional morphology, while declarative linguistic rules handle irregular and constraint-sensitive phenomena that cannot be reliably captured through finite-state representations alone. The interaction between the two components is governed by a deterministic integration and conflict resolution strategy, ensuring predictable and interpretable analysis. Experimental evaluation demonstrates that the hybrid approach achieves higher coverage and accuracy than finite-state-only and rule-based-only baselines. The results confirm that a generator-based hybrid architecture provides an effective and scalable solution for mor- phological analysis in low-resource linguistic settings.

🏷️ Keywords

Morphological Analyzer Generator Finite-State Morphology Rule-Based Morphology Low-Resource Languages

🔗 DOI

View DOI - (https://doi.org/10.36713/epra25947)

📚 How to Cite:

Alok Kumar, Manisha Kumari Deep , A HYBRID FINITE-STATE AND RULE-BASED ARCHITECTURE FOR THE ANGIKA MORPHOLOGICAL ANALYZER GENERATOR , Volume 11 , Issue 1, January 2026, EPRA International Journal of Research & Development (IJRD) , DOI: https://doi.org/10.36713/epra25947

🔗 PDF URL

https://cdn.epratrustpublishing.com/article/202602-02-025947.pdf

📄 PDF Preview

Click the button above to load the PDF.